It strikes me that one of the big differences between the 1.0 class of data aggregators and the 2.0 class is the difference between "back office" and "live" applications. The credit card company mines its database to select you for direct mail offers; it may even get close to real time in monitoring your card activity for fraud or credit limit detection. But Google or Amazon mines its database in real time and builds the results right into its customer-facing applications.
How can you turn your data from a back-office asset that benefits only your company into a driver of dynamic user-facing services? How can you make your software more responsive to its users, harnessing their collective intelligence and learning from them? How can you make your software or site more alive?
Inspiration: Tim O'Reilly